Wednesday, December 5, 2012

Breaking Silence

So much for that promise of weekly updates, eh?

There has been a reason for that. Most of my attention has (had) shifted to a separate project related to course material for our Forecasting classes. I had to take some time to work on that project while students were (are) using it.

That project now seems to be more or less out of the way. As such I now can devote my attention back to some other projects that have been in the works.

The grand "re-working" of our analysis products is still in progress. There are a few hurdles to jump through with that. I'm not out to make any promises about when that will all be done... beyond saying that our hope is to have it all done by the end of the year... but who knows, that might be a bit of an under-estimate. We may take until mid-January to finish that stuff up.

In the mean time I plan on working on a couple other "minor" projects. I will (finally) be creating a way for mobile users to switch between the mobile catered "JSAni" satellite pages, to the desktop catered "Flanis" pages. That is mainly a fix for strict Android users running a version of Gingerbread that may not properly load overlays.

Speaking of our satellite data; Since GOES 14 has shifted it's orbit into the position of GOES 13, and has now permanently taken over the role as GOES East, a strange thing has occurred. I started to notice a familiar pattern return to the diurnal change in filesize for our satellite image. Short story version of that; I'm re-introducing "night-time removal" of images to our visible satellite loops. It's not perfect, however. It only seems to work with sectors that are covered almost entirely by GOES East (a.k.a. GOES 14). Many of our sectors along the WRN CONUS will not remove night-time images from loops. But I feel something is better than nothing in this case.

I also have a whole host of random (but COOL!) ideas to inject into our model data. Apart from purely cleaning up code (!), I'm working on something that should help our model comparison products work better. Without going into too much detail, I'm hoping to create a more seamless experience with using a comparison product, and then looking at model products for a given model/run/sector/hour. In addition, I've already started toying with some "run status" type data. Meaning you, the user, should have more of a cue when new model runs are coming in, and just how "done" they are. Mike and Victor have also been plotting and scheming some improvements to the images themselves. Perhaps even some new products. Last but not least... Improved sector coverage! Hopefully we will delight some west coast users.

That's all for now. Back to work. Please pardon any typos or grammatical errors in this post.